368 EXPERIMENTAL STUDY OF INHERITANCE 



quently occur among plants and give rise to true-breeding 

 varieties (see Chapter III.). 



Now it is evident that, if Mendel's Law applies in such cases, 

 the mutation, once present, is not likely to be lost or swamped 

 by inbreeding with the normal types. Thus, through Mendel's 

 discovery we are led to a new view of organic evolution, in which 

 we attach less importance to the minute fluctuations on which 

 Darwin relied, and more importance to mutations or saltatory 

 variations. 



Light thrown on Variation. — Mendelian experimentation hat. 

 thrown light on at least some kinds of variation. In connection 

 with the colours of flowers and of the coats of mammals, it has 

 been shown that varieties may arise by the loss or modification 

 of unit-characters. Thus in th£ case of a rabbit, some colour- 

 factor may drop out altogether, giving albinos, or the pattern- 

 factor of the individual hairs may drop out, giving a mingling 

 of pigment which appears black, or the factor for black may 

 drop out, giving brown and cinnamon varieties. But what 

 does this " dropping out " mean ? Prof. Castle answers : " Loss 

 of a unit-character might easily come about by an irregular 

 cell-division in which the material basis of a character failed to 

 split, as normally. . . . On the other hand, a modified condition 

 of a unit-character might possibly result from unequal division 

 of the material basis of a character, so that one of the cell- 

 products would transmit the character in weakened intensity, 

 the other in increased intensity " (1911, p. 86). 



Some Mendelians would also admit the idea that a unit- 

 character may lose some of its " potency," some of its power of 

 " dominating " or of asserting itself in development — just as 

 it might on Weismann's theory of germinal selection. We have 

 already referred to the not infrequent imperfection of the domin- 

 ance of a dominant character. It may be that this is due to a 

 weakening of the potency of a particular unit-character ; it may 

 be that something must be allowed for the condition of the 



